AI Is Wrecking an Already Fragile Job Market for College Graduates

Companies have long leaned on entry-level workers to do grunt work that doubles as on-the-job training. Now ChatGPT and other bots can do many of those chores.

What do you hire a 22-year-old college graduate for these days?

For a growing number of bosses, the answer is not much—AI can do the work instead.

At Chicago recruiting firm Hirewell, marketing agency clients have all but stopped requesting entry-level staff—young grads once in high demand but whose work is now a “home run” for AI, the firm’s chief growth officer said. Dating app Grindr is hiring more seasoned engineers, forgoing some junior coders straight out of school, and CEO George Arison said companies are “going to need less and less people at the bottom.”

Bill Balderaz, CEO of Columbus-based consulting firm Futurety, said he decided not to hire a summer intern this year, opting to run social-media copy through ChatGPT instead.

Balderaz has urged his own kids to focus on jobs that require people skills and can’t easily be automated. One is becoming a police officer.

Having a good job “guaranteed” after college, he said, “I don’t think that’s an absolute truth today any more.”

There’s long been an unwritten covenant between companies and new graduates: Entry-level employees, young and hungry, are willing to work hard for lower pay. Employers, in turn, provide training and experience to give young professionals a foothold in the job market, seeding the workforce of tomorrow.

A yearslong white-collar hiring slump and recession worries have weakened that contract. Artificial intelligence now threatens to break it completely.

That is ominous for college graduates looking for starter jobs, but also potentially a fundamental realignment in how the workforce is structured. As companies hire and train fewer young people, they may also be shrinking the pool of workers that will be ready to take on more responsibility in five or 10 years. Companies say they are already rethinking how to develop the next generation of talent.

AI is accelerating trends that were already under way.

With each new class after 2020, an ever-smaller share of graduates is landing jobs that require a bachelor’s degree, according to a Burning Glass Institute analysis of labor data. That’s happening across majors, from visual arts to engineering and mathematics. And unemployment among recent college graduates is now rising faster than for young adults with just high-school or associate degrees.

Meanwhile, the sectors where graduate hiring has slowed the most—like information, finance, insurance and technical services—are still growing, a sign employers are becoming more efficient and see no immediate downside to hiring fewer inexperienced workers, said Matt Sigelman, Burning Glass’s president.

“This is a more tectonic shift in the way employers are hiring,” Sigelman said. “Employers are significantly more likely to be letting go of their workers at the entry level—and in many cases are stepping up their hiring of more experienced professionals.”

After dancing around the issue in the 2½ years since ChatGPT’s release upended the way almost all companies plan for their futures, CEOs are now talking openly about AI’s immense capabilities likely leading to deep job cuts. Top executives at industry giants including Amazon and JPMorgan have said in recent weeks that they expect their workforces to shrink considerably. Ford CEO Jim Farley said he expects AI will replace half of the white-collar workforce in the U.S.

For new graduates, this means not only are they competing for fewer slots but they are also increasingly up against junior workers who have been recently laid off.

While many bosses say they remain committed to entry-level workers and understand their value, the data is increasingly stark: The overall national unemployment rate is at about 4%, but for new college graduates, it was 6.6% over the past 12 months ending in May.

At large tech companies, which power much of the U.S. economy, the trend is perhaps more extreme.

Venture-capital firm SignalFire found that among the 15 largest tech companies by market capitalization, the share of entry-level hires relative to total new hires has fallen by 50% since 2019. Recent graduates accounted for just 7% of new hires in 2024, down from 11% in 2022. A May report by the firm pointed to shrinking teams, fewer programs for new graduates and the growing influence of AI.

Jadin Tate studied informatics at the University at Albany, hoping to land a job focused on improving the user experience of apps or websites. The week before graduation, his mentor leveled with him: That field is being taken over by AI. He warned it may not exist in five years.

Tate has attended four conventions this year, networking with companies and asking if they are hiring. He has also applied to dozens of jobs, without success. Several of his college friends are working retail and food-service jobs as they apply for white-collar roles or before their start dates.

“It has been intimidating,” Tate said of his job search.

Indeed, recent graduates and students are fighting over a smaller number of positions geared at entry-level workers. There were 15% fewer job postings to the entry-level job-search platform Handshake this school year than last, while the number of applications per job rose 30%, according to the platform. Internship postings and applications saw similar trend lines between 2023 and 2025.

Less gruntwork, more mentoring

The shift to AI presents huge risks to companies on skill development, even as they enjoy increased efficiency and productivity from fewer workers, said Chris Ernst, chief learning officer at the HR and finance software company Workday.

Ernst said his research shows that workers mostly learn through experience, and then the remainder comes from relationships and development. When AI can produce in seconds a report that previously would have taken a young employee days or weeks—teaching that person critical skills along the way—companies will have to learn to train that person differently.

“Genuine learning, growth, adaptation—it comes from doing the hard work,” he said. “It’s those moments of challenge, of hardship—that’s the crucible where people grow, they change, they learn most profoundly.”

Among other things, Ernst said employers must be intentional about connecting young workers with colleagues and making time to mentor them.

At the pipeline operator Williams, based in Tulsa, Okla., the company realized that thanks to AI young professionals were performing less of the drudgework like digging into corporate data that historically has taught them the core of the business.

The company this year started a two-day onboarding program where veteran executives teach new hires the business fundamentals. Chief Human Resources Officer Debbie Pickle said that increased training will help new hires develop without loading them down with gruntwork.

“These are really bright, top talent people,” she said. “We shouldn’t put a cap on how we think they can add value for the company.”

Still, Pickle said, the increased efficiency will allow the company to expand the business while keeping head count flat in the future.

Some of the entry-level jobs most at risk are the most lucrative for recent graduates, including on Wall Street and in big law firms where six-figure starting salaries are the norm. But those jobs have also been famously menial for the first few years—until AI came along.

The investment firm Carlyle now pitches to prospective hires that they won’t be doing grunt work. Junior hires go through AI training and a program called “AI University” in which employees share best practices and participate in pilot programs, said Lúcia Soares, the firm’s chief information officer.

In the past, she said, junior hires evaluating a deal would find articles on Google, request documents from companies, review that information manually, highlight details and copy and paste information from one document to another. Now, AI tools can do almost all of that.

“That analyst still has to go in and make sure the analysis is accurate, question it, challenge it,” she said. “The nature of the brain work that needs to go into it is very much the same. It’s just the speed at which these analysts can move.”

She said Carlyle has maintained the same volume of entry-level hiring but said 90% of its staff has adopted generative AI tools that automate some work.

‘Messy transition’

Carlyle’s reliance on young staff to check AI’s output highlights what many users know to be true: it still struggles in some cases to do the work of humans effectively. Still, many executives expect that gap to close quickly.

At the New York venture-capital firm Primary Venture Partners, Rebecca Price said she’s encouraging CEOs of the firm’s 100 portfolio companies to think hard about every hire and whether the role could be automated.

She said it’s not that there are no entry-level jobs, but that there’s a gap between the skills companies expect out of their junior hires in the age of AI and what most new graduates are equipped with out of school. An engineer in a first job used to need basic coding abilities: now that same engineer needs to be able to detect vulnerabilities and have the judgment to determine what can be trusted from the AI models.

New grads must also learn faster and think critically, she said—skills that many of the newest computer-science grads don’t have yet.

“We’re in this messy transition,” said Price, a partner at the firm. “The bar is higher and the system hasn’t caught up.”

Students are seeing the transition in real time. Arjun Dabir, a 20-year-old applied math major at the University of California, Irvine, said when he applied for internships last year, companies asked for knowledge of coding languages. Now, they want candidates who are familiar with how AI “agents” can automate certain tasks on behalf of humans—or “agentic workflows” in the new vernacular.

“What is an intern going to do?” Dabir said as drones buzzed overhead nearby at an artificial intelligence convention in June in Washington, DC. The work typically done by interns, “that task is no longer necessary. You don’t need to hire someone to do it.”

Venture capitalist Allison Baum Gates said young professionals will need to be more entrepreneurial and gain experience on their own without the standard track of starting as an analyst or a paralegal and working their way up. Her firm, SemperVirens, invests in healthcare startups, workforce technology companies and fintech firms, some of which are replacing entry-level jobs.

“Maybe I’m wrong and this leads to a wealth of new jobs and opportunities and that would be a great situation,” she said. “But it would be far worse to assume that there’s no adverse impact and then be caught without a solution.”

Rosalia Burr, 25, is trying to avoid such an outcome. She graduated in 2022 and quickly joined Liberty Mutual Insurance, where she had interned twice during college at Arizona State University. She was laid off from her payroll job in December.

Running has soothed her anxiety. This spring, however, she tore her hip flexor and had to rest to heal. Job rejections, as she was stuck inside, hit extra hard. “I felt that I was failing.”

Her goal now is to find a client-facing job. “If you’re in a business back-end role, you’re more of a liability of getting laid off, or your job being automated,” she said. “If you’re client facing, that’s something people can’t really replicate” with AI.

(http://www.autoadmit.com/thread.php?thread_id=5756016&forum_id=2\u0026hid=#49140612)

...what was it cr for before?

(http://www.autoadmit.com/thread.php?thread_id=5756016&forum_id=2\u0026hid=#49141637)

This plus offshoring plus H1Bs will end up screwing over a lot of people.

(http://www.autoadmit.com/thread.php?thread_id=5756016&forum_id=2\u0026hid=#49141167)

not only are lawyers some of the most makework bullshit jobs out there, but candy ride is a poor, unsuccessful solo. so not only does he work a job that's fake fundementally but it's also fake specifically since he doesn't make a livable income doing it.

(http://www.autoadmit.com/thread.php?thread_id=5756016&forum_id=2\u0026hid=#49142761)

I checked the code and cases it cited to confirm everything was legit. Took like five minutes not flame.

In another case, high conflict family law case, mother’s lawyer lodged a 54 page PDF that had all these school and medical records including an IEP and a IPP.

I started looking through that shit and said fuck this, loaded it up and said, “Anything in here that helps my client?”

Of course I typed a little summary of the case. It instantly told me that on page 34, mother makes an admission that helps me.

It summarized all the arguments I could make from those documents, categorized them, cited to the PDF page number and gave me quotes. Instantly.

Sometimes I write these super long emails that take me .7 to an hour and AI will polish it up.

I’ve asked it to compare two documents and note any substantive changes. After reviewing manually first, of course.

My favorite new use is when opposing counsel is being a dick and sends me a shitty email I’ll ask it to draft me a “patronizing but polite” response. Then I slightly edit it to remove any obvious AI shit.

(http://www.autoadmit.com/thread.php?thread_id=5756016&forum_id=2\u0026hid=#49142046)

Buc-ees manager

(http://www.autoadmit.com/thread.php?thread_id=5756016&forum_id=2\u0026hid=#49141976)